A population-based study of the risk of osteoporosis and fracture with dutasteride and finasteride

Osteoporosis is an under-appreciated cause of morbidity and mortality in men [1, 2]. Globally, one in three osteoporotic hip fractures occur in men, and a higher proportion of men than women die in the first year following a hip fracture, with mortality rates of 37.5 and 28.2%, respectively [3]. Evidence suggests that androgen deficiency contributes to bone loss and fracture risk in older men [4‐7]. The observation that men with osteoporosis have lower dihydrotestosterone concentrations than men with normal bone mineral density suggests an important role of this androgen in bone homeostasis [8, 9]. Dihydrotestosterone is approximately ten times more potent than its precursor testosterone, and is the preferred ligand for androgen receptor transactivation [10]. Because 5-alpha reductases convert testosterone to dihydrotestosterone, inhibitors of these enzymes could conceivably predispose older men to osteoporosis and fractures [11].

Dutasteride and finasteride are 5-alpha reductase inhibitors that are equally effective treatments for benign prostatic hyperplasia [12]. Although considered clinically interchangeable, the two drugs differ in their spectrum of 5-alpha reductase inhibition. Finasteride is a selective inhibitor of the type 2 isoform of 5-alpha reductase, which is found predominantly in the prostate, while dutasteride additionally inhibits the more widespread type 1 isoform, which is the predominant 5-alpha reductase in osteoblasts [11‐13]. Because dutasteride is both a more potent and less selective inhibitor of 5-alpha reductases, it reduces circulating serum dihydrotestosterone by 90 to 95%, compared with 60 to 70% for finasteride [11, 14, 15] In addition, dutasteride mediated-inhibition of 5-alpha reductase in osteoblasts could conceivably suppress local production of dihydrotestosterone [13]. These effects may have implications for bone health, particularly among older men, because bone loss accelerates rapidly after the age of 70 years [1]. However, whether dutasteride imparts a higher risk of osteoporosis and fractures in older men receiving 5-alpha reductase inhibitors is unknown.

Several studies have examined the effects of 5-alpha reductase inhibitors on bone mineral density. Although no significant changes in bone mineral density were observed in a one-year randomized trial of dutasteride and finasteride, the study was small and restricted to men aged 18 to 55 years [16]. Similar findings were observed in a small non-randomized study of men aged 60 to 78 years who were followed for up to two years after treatment initiation [17]. Observational studies have yielded inconsistent findings, ranging from no association between 5-alpha reductase inhibitors and bone disease to both a higher and lower risk imparted by these drugs [18‐22]. Importantly, no observational study has specifically explored whether the available 5-alpha reductase inhibitors carry differential risks of osteoporosis or fracture in older men. This is important because benign prostatic hyperplasia is common in older men, 5-alpha reductase inhibitors are commonly prescribed for this indication, and male osteoporosis imparts a substantial burden on both health and society. We compared the risk of osteoporosis and fractures in older men commencing treatment with either dutasteride or finasteride. We hypothesized that, by virtue of more pronounced and widespread 5-alpha reductase inhibition, dutasteride might be associated with a heightened risk of these outcomes.

During the six-year accrual period, we identified 31,615 individuals treated with dutasteride who were matched to an equal number of subjects treated with finasteride. Patients treated with dutasteride were followed for a median of 328 days (interquartile range 74 to 730 days), while those treated with finasteride were followed for a median of 313 days (interquartile range 64 to 730 days). Collectively, individuals in the cohort contributed a total of 65,804 person-years of follow-up. After propensity score matching, the two groups were highly similar in terms of demographics, medical illnesses and concomitant medications associated with osteoporosis and fractures (Table 1). However, patients differed with respect to the specialty of prescribing physician, with dutasteride-treated individuals more likely to have treatment prescribed by a urologist [16,862 (53.3%) versus 11,402 (36.1%)], whereas finasteride was more likely to be initiated by family physicians [13,553 (42.9%) versus 8924 (28.2%)] (Table 1).

In the primary analysis, osteoporosis was diagnosed in 1569 individuals, with a lower rate observed among men prescribed dutasteride relative to finasteride [2.2 versus 2.6 per 100 person-years; hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.72 to 0.93]. This difference was no longer statistically significant following adjustment for physician specialty (adjusted HR 0.90, 95% CI 0.78 to 1.02) (Table 2). In the secondary analysis, we found no difference in the risk of osteoporotic or fragility fracture between men prescribed dutasteride or finasteride (adjusted HR 1.04, 95% CI 0.86 to 1.25) (Table 2). As expected, we found no difference in cataract surgery between the two groups (HR 0.98; 95% CI 0.89 to 1.09).

In this population-based study, we found no difference in the risk of fractures between older men treated with dutasteride or finasteride. Conversely, and in contrast to our study hypothesis, we found a lower incidence of osteoporosis in dutasteride-treated men compared with men treated with finasteride, but this was not statistically significant after adjustment for physician specialty. Our findings do not support a heightened risk of adverse bone outcomes among older men treated with dutasteride relative to finasteride, and that this drug can be used safely in this regard.

The finding that dutasteride did not increase the risk of adverse bone outcomes was unexpected in light of evidence that it inhibits dihydrotestosterone production to a greater extent than finasteride and that men with osteoporosis have lower levels of this androgen than men with normal bone mineral density [9, 15]. One possible explanation for this discordance is that dutasteride-mediated inhibition of osteoblast 5-alpha reductase activity is compensated for by higher levels of intracellular testosterone, which undergoes subsequent aromatization to estradiol [6, 13, 33]. This reasoning is supported by studies in men demonstrating better correlation between bone mineral density and serum estradiol rather than testosterone [33, 34]. Because it does not inhibit 5-alpha reductase in osteoblasts, this phenomenon would not be expected to occur with finasteride.

Another potential explanation for the lower crude rates of osteoporosis among dutasteride-treated men relates to the greater frequency with which this drug was prescribed by urologists relative to finasteride. Despite universal coverage of physician services, Ontario residents with high educational attainment have more contact with specialists and bypass primary care to reach specialists more often than those with lower education [35]. Extending these findings to our study, it is possible that the lower incidence of osteoporosis among dutasteride recipients reflects in part greater knowledge about osteoporosis prevention and a healthy user effect, an assertion supported by the loss of a statistically significant association between dutasteride and osteoporosis after adjustment for physician specialty. Detection bias may also account for this finding because osteoporosis screening is more likely to be undertaken by family physicians than urologists, with the former ordering approximately 80% of bone mineral density tests in Ontario [36]. Importantly, however, rates of osteoporosis screening among men at high risk for the disease consistently lag far below those of women. In Ontario, age-standardized rates of bone mineral density testing within 6 months of a fracture were 8 per 100 men, compared with 16.1 per 100 women, in 2009/2010 [37]. Similarly, only 11% of Ontario men between the ages of 68 to 70 who had never previously undergone bone mineral testing received such a test in 2009/2010, compared with over 40% of Ontario women in this same high-risk group [37]. These findings are supported by those of a Canadian cohort study in which 90% of men with fragility fractures remained undiagnosed and untreated for osteoporosis despite participation in a five-year study with the potential to raise awareness of the disease among participants and their physicians [38]. Interventions to promote osteoporosis diagnosis and treatment among men are warranted.

Our study builds on previous observational studies examining the association between adverse bone outcomes and 5-alpha reductase inhibitors in several ways. First, ours is the first study to specifically compare the risk of osteoporosis and fractures between dutasteride and finasteride. Previous studies have risk estimates for 5-alpha reductase inhibitor exposure relative to a control group of untreated men [18‐22]. Second, we generated additional data regarding the risk of osteoporosis in men treated with 5-alpha reductase inhibitors, an outcome examined in only one case control study with 47 osteoporosis patients exposed to one of these drugs [21]. Finally, we were able to demonstrate the important confounding role of physician specialty in the association between adverse bone outcomes and 5-alpha reductase inhibitors. Specifically, the finding that adjustment for physician specialty reversed the association between dutasteride and osteoporosis suggests that future studies of these drugs consider accounting for health service delivery and other indicators of a potential healthy user effect.

Our study has some limitations. We used administrative data rather than measurement of bone mineral density to identify patients with osteoporosis, and codes used for ascertaining osteoporosis have not been validated among older men. We had no access to relevant clinical variables such as smoking history, alcohol use, body mass index and use of over the counter calcium and vitamin D supplementation. However, these limitations apply equally to both groups of patients. We followed patients for two years from the time of drug initiation; it is possible that additional outcomes may have been observed with longer follow-up. We selected a two-year follow-up because of the advanced age of our cohort, reasoning that prostate cancer and frailty could emerge as important confounders with a longer follow-up period. In addition, we had no information regarding medication adherence. Finally, because we conducted our analyses in men aged 66 years and over with no prior history of osteoporosis or fractures, our findings may not be applicable to younger patients and those with pre-existing bone disease.

We found no difference in the incidence of a new diagnosis of osteoporosis among older men treated with either dutasteride or finasteride. Similarly, we found no differential risk of fractures with dutasteride. These findings provide a measure of reassurance that widespread inhibition of 5-alpha-reductase does not negatively affect bone health in older men, and that dutasteride can be safely used in this regard.